Productivity Analysis of Electricity Distribution Commerce - - PowerPoint PPT Presentation

Productivity Analysis of Electricity Distribution

Commerce Commission Workshop 2 May 2014 Denis Lawrence

www.economicinsights.com.au 2

Workshop Outline

• Background
• Where productivity measures fit within EDB regulation
• What is TFP and partial productivity?
• How do we measure productivity?
• Previous productivity studies
• Output specification issues
• Functional vs billed outputs
• Proposed approaches
• Morning tea
• Input specification issues
• Opex
• Capital
• Proposed approaches

www.economicinsights.com.au 3

BACKGROUND

www.economicinsights.com.au 4

Regulatory background

• NZ EDB regulation has moved from being purely productivity-

based regime under the former thresholds regime to a high level building blocks approach for non-exempt EDBs

• Under building blocks, annual price caps are set to equate the

present value of forecast costs with the present value of forecast revenue for each regulated EDB

• 17 out of the 29 NZ EDBs are currently non-exempt and

subject to a default price-quality path under Part 4 of the Commerce Act

• Commerce Commission will reset the default price path for the

non-exempt EDBs effective from 1 April 2015

www.economicinsights.com.au 5

Regulatory background (2)

• Reset will involve specifying maximum starting prices for each

regulated EDB and a common ‘rate of change’ of those prices across all regulated EDBs

• Starting prices will be set to equate present value of forecast

costs with present value of forecast revenue for each non– exempt EDB, subject to a common rate of change

• Rate of change must be based on the long–run average

productivity improvement rate achieved by either or both of EDBs in New Zealand, and suppliers in other comparable countries, using appropriate productivity measures

www.economicinsights.com.au 6

Regulatory background (3)

• If starting prices are based on the current and projected

profitability of each supplier, the rate of change will not affect the amount of revenue the individual EDB can expect to recover over the regulatory period.

• This is because starting prices for each regulated EDB would

simply be adjusted to offset any alteration to the common rate

• f change to maintain the equality between the present value
• f expected revenues and the present value of expected costs

for that EDB over the regulatory period.

• This means the regulatory outcome for each EDB is not

affected by the measured long–run average productivity improvement rate used to set the rate of change of prices

www.economicinsights.com.au 7

Regulatory background (4)

• But the forecast partial opex and capital productivities can

impact the level of forecast costs and therefore the present value of allowable revenue over the regulatory period for each EDB.

• In the case of opex, forecast opex is set using the following:

www.economicinsights.com.au 8

• TFP is the ratio of the quantity of total output produced to the

quantity of total inputs used

• TFP growth between two periods is given by:

TFP Growth = (Q2/X2)/(Q1/X1) = (Q2/Q1) / (X2/X1) where Q is total output, X is total input and 1,2 are years

• How do we form measures of total output and input

quantities?

• We use index number methods which aggregate diverse
• utput quantities (weighted by revenue or output cost shares)

and input quantities (weighted by cost shares) into indexes

• A simple way of viewing TFP growth is as the ratio of the

weighted average change in output quantities to the weighted average change in input quantities

• Requires values and quantities for all outputs and inputs

Total factor productivity

www.economicinsights.com.au 9

• Partial productivity growth is the ratio of the weighted

average change in output quantities to the change in the quantity of a particular input

• The key partial productivities in this instance are opex partial

productivity and capital partial productivity

Key issues:

• How do we measure EDB outputs?
• Is it what the EDB charges for or is it what it is expected

to provide?

• How do we measure EDB inputs?
• Main issue is how we measure the quantity of annual

capital input?

Partial productivity

www.economicinsights.com.au 10

• Recent resets of the default price path have been based on the

results of studies undertaken in 2009 by Economic Insights (for the Commission) and Pacific Economics Group (for the ENA)

• Despite using different approaches and specifications both

studies produced similar electricity distribution industry productivity results

• Economic Insights found an annual TFP growth rate of just over

1 per cent for the 13 years to 2008, around the same as for the economy as a whole over the same period

• PEG found an annual TFP growth rate of 1.2 per cent for the 10

years to 2008

• Based on these findings the Commission set an X factor of

0 per cent (leading to a rate of change of CPI-0 per cent)

• 0 per cent also assumed for opex partial productivity growth

Previous productivity studies

www.economicinsights.com.au 11

Previous productivity studies (2)

0.9 1.0 1.1 1.2 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Economic Insights EDB TFP PEG EDB TFP SNZ MFP

www.economicinsights.com.au 12

Previous productivity studies (3)

0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Economic Insights Opex PFP PEG Opex PFP

www.economicinsights.com.au 13

Economic Insights (2009) EDB TFP specification

3 outputs:

• throughput in GWh
• customer numbers
• system capacity based on MVA*kms

Outputs were output cost share weighted 4 inputs

• pex
• verhead lines in MVA–kms
• underground lines in MVA–kms
• transformers in KVA and other capital

Inputs weights formed using exogenous capital cost

www.economicinsights.com.au 14

PEG (2009) EDB TFP specification

3 outputs:

• customer numbers
• throughput in GWh
• peak demand as measured by the non–coincident peak

in GW Outputs were revenue share weighted 2 inputs:

• pex
• capital (measured by constant price depreciated asset

value) Inputs weights formed using endogenous capital cost

www.economicinsights.com.au 15

OUTPUTS

www.economicinsights.com.au 16

• Outputs in a TFP study should cover the main things the

EDB produces which are valued by customers

• In a competitive market the prices charged by a firm will

reflect its costs of producing key dimensions valued by customers

• This makes it easy to identify and measure a firm’s output
• But in a natural monopoly there may be a divergence

between what firms charge for, what their main cost drivers are and what their customers (and regulators) value

• This makes it harder both to identify and to measure a

network business’s outputs that should be included in TFP measures

Output coverage

www.economicinsights.com.au 17

• Billed outputs are those items an EDB actually charges

customers for

• EDB charging practices have typically evolved on an ease of

implementation basis rather than on a network cost- reflective basis

• High proportion of charges are often on energy throughput

but throughput is a minor driver of EDB costs

• Dimensions that customers value such as availability of

supply, reliability, speedy restoration after any interruption are not explicitly charged for but may be significant cost drivers

• Functional outputs are all those services EDBs provide to

customers which are valued by customers (of which billed

• utputs are a subset)

Billed vs functional outputs (1)

www.economicinsights.com.au 18

• Under building blocks the revenue requirement is set based
• n the EDB being expected to meet a range of performance

standards and other deliverables

• Prices then have to be consistent with broad regulatory

pricing principles

• In the case of building blocks, it will be important to measure
• utput in a way that is broadly consistent with the output

dimensions implicit in the setting of EDB revenue requirements

• These output dimensions may be different to the outputs the

EDB charges its customers for

• Points to functional rather than billed outputs specification in

building blocks but sensitivity analysis is important

Billed vs functional outputs (2)

www.economicinsights.com.au 19

• Energy delivered

Bulk of charges but little impact on costs Proxy for load? Road analogy

• Customer numbers

Proxy for fixed output components Need to disaggregate?

• Demand-based outputs

Industrial customers usually charged on basis of actual peak demand

• r reserved capacity

Data available but not easily aggregated System proxies problematic

Billed outputs

www.economicinsights.com.au 20

• System capacity

Captures capacity to deliver (road analogy) Line capacity or line and transformer capacity? How can DSM and embedded generation be allowed for?

• Peak demand

Important cost driver but customers see reliability at peak as relevant

• utput

Volatility and incentive effects

• Reliability

Important to customers Some practical issues to be resolved Should other aspects of service quality be included?

Other functional outputs

www.economicinsights.com.au 21

What we plan to look at

1) Functional Outputs Quantity Value Price Customers (No) Capacity (kVA*kms) Throughput (GWh) Revenue * Cost share Revenue * Cost share Revenue * Cost share Value / Customers Value / kVA*kms Value / GWh Value / Customers Value / MVA Value / GWh Revenue Revenue Revenue Customers (No) Max demand (MW) Throughput (GWh) Price Value Quantity 2) Billed outputs

www.economicinsights.com.au 22

Output quantities from 2009 studies

0.8 0.9 1.0 1.1 1.2 1.3 1.4 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Customer Numbers Maximum Demand System Capacity Energy

www.economicinsights.com.au 23

Issues for discussion

• Availability of information on billed output quantities and

revenues and ability to aggregate into a usable form

• Is system maximum demand a reasonable proxy for the

sum of maximum demands for customers that are charged

• n that basis?
• Have revenue shares changed significantly in the past

decade?

• Should system capacity take account of both transformer

and line/cable capacity?

• Is it reasonable to include throughput as an EDB output?
• Should reliability be included? If so, how?
• Other questions/issues?

www.economicinsights.com.au 24

INPUTS

www.economicinsights.com.au 25

Different types of inputs

• Opex (labour, materials and services) and capital (lines,

cables and transformers)

• Different treatment of these inputs required in productivity

studies

• Obtaining price and quantity of non-durable opex inputs is

relatively straightforward

• But obtaining accurate annual cost and annual input quantity
• f durable capital inputs is more challenging
• Capital service flow is the quantity of capital inputs’

contribution to production each year

• Annual user cost is cost of using capital inputs each year

www.economicinsights.com.au 26

Opex coverage

• Separate data on labour, materials and services

generally not available

• Given extent of contracting out, generally use aggregate
• pex
• Diverse composition requires deflating opex by a price

index to derive the quantity of opex inputs indirectly

• Need uniform treatment of asset refurbishment and

allocation of corporate overheads

• Have previously used scaled up Direct line costs per

kilometre and Indirect line costs per customer to ensure

• nly current nondurable input use is included
• Propose to roll this forward using change in opex from

new IDD after 2008

www.economicinsights.com.au 27

Opex price index

• Around 60% of opex is typically made up of labour costs
• Remainder covers operational, office and professional

services and consumables

• EDB-specific price indexes generally not available
• Use SNZ economy-wide or broad EGWW sectoral price

indexes instead

• Use Labour price index for labour price (and price for all
• f opex in previous studies)
• Could use Producer price index for non-labour component
• f opex

www.economicinsights.com.au 28

Opex specification

Quantity Cost Price Nominal opex / Weighted average price index Opex (scaled up from Direct and Indirect costs and rolled forward using total opex after 2008) Weighted average of an SNZ labour price index producer price index Advantages

• Uses readily available data
• Unlikely to be any practical alternatives

Limitations

• Dependent on changes in SNZ price indexes accurately reflecting changes

in opex prices faced by all EDBs

• Assumes all EDBs face the same levels of opex component prices
• Assumes the same range of opex coverage and functions for all EDBs

www.economicinsights.com.au 29

Capital input quantities

• Capital service flow is not directly observable – have to

use proxy measures instead

• No proxy measure will be perfect – all have pros and

cons

• Best proxy measure will depend on:

Physical depreciation profile of constituent assets Robustness of the data used to form the proxy – the

more accurate the better and the fewer assumptions that have to be made, the better

www.economicinsights.com.au 30

Physical depreciation profiles

• Important to distinguish between physical and financial

regulatory depreciation

• Asset physical ‘carrying capacities’ can have one of four

different physical depreciation profiles:

• ne hoss shay (no reduction in carrying capacity

throughout the asset’s life)

hyperbolic (small decline in early years, larger decline

in later years of the asset’s life)

straight–line (ie equal absolute reduction each year of

the asset’s life), or

geometric (equal percentage decline over lifetime, ie

bigger absolute falls in early years and smaller absolute falls in later years)

www.economicinsights.com.au 31

EDB asset physical depreciation

• EDB assets subject to little physical deterioration over

their lifetime provided they are properly maintained

• Failure rates may be higher towards the end of asset life

leading to Weibull distribution of asset ages

• But service flow relatively constant over the asset’s life

meaning one hoss shay depreciation is closest

• Can approximate service flow under one hoss shay by

using physical measures (eg MVA-kms) or gross capital stock in constant prices

• Using constant price depreciated asset value proxies

assumes depreciation profile other than one hoss shay

• If the decay of the asset over time is overestimated by the

proxy then productivity growth will also be overestimated

www.economicinsights.com.au 32

• Annual user cost of capital consists of:

depreciation plus opportunity cost minus capital gains

• Equivalent building block concepts are return of capital

(depreciation less inflation allowance) and the return on capital (opportunity cost)

• Building blocks also implicitly includes concept of ex-ante

financial capital maintenance based on RAB and WACC

• Preferable to have (approximate) consistency between the

ways annual capital costs are calculated in building blocks and productivity measurement

Annual user cost of capital

www.economicinsights.com.au 33

Alternative capital specifications (1)

Quantity Annual user cost Price O/H MVA-kms U/G MVA-kms Transformers & other KVA Return of & on O/H capital Return of & on U/G capital Return of & on Trf & other capital O/H AUC/MVA-kms U/G AUC/MVA-kms Trf & other AUC/KVA

Physical proxies

Advantages

• Reflects individual component carrying capacities
• Uses the most robust data available and captures actual asset lives
• Approximates productive capital stock used by leading statistical agencies

for aggregate structures

• Ensures broad consistency with building blocks calculations

Limitations

• Small amount of extra data required

www.economicinsights.com.au 34

Alternative capital specifications (2)

Quantity Annual user cost Price Constant price depreciated asset value Revenue minus opex AUC/Const price depreciated RAB

Depreciated asset value proxy

Advantages

• Easy to implement
• Uses existing regulatory data

Limitations

• Unlikely to reflect carrying capacities of assets
• Dependent on accuracy and consistency of initial capital base values
• Assumes consistency of RAB treatment over time and across EDBs
• May not capture actual asset lives
• Dependent on SNZ CGPI accurately capturing prices paid by EDBs
• Unlikely to be consistent with financial capital maintenance

www.economicinsights.com.au 35

Input quantities from 2009 studies

0.7 0.8 0.9 1.0 1.1 1.2 1.3 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Economic Insights Opex PEG Opex Overhead Underground Transformers PEG Capital

www.economicinsights.com.au 36

Issues for discussion

• Is scaling up Direct and Indirect costs from the old IDD

and rolling forward using total opex from new IDD the best way to form the measure of annual nondurable input use?

• How representative are the high level price indexes of the

price changes EDBs actually face?

• Which profile most accurately reflects the depreciation of

EDB physical capital ‘carrying capacity’?

• Which data source – financial or physical – gives the most

robust reflection of EDB capital annual input quantities?

• How important is it for the capital weight used to reflect the

true cost of capital rather than the realised cost?

• Other questions/issues?